Joint Modeling of Survival Time with Another Outcome in Clinical Trials or Observational Studies

• Activity Number: 107
• Type: Invited
• Date/Time: Monday, August 1, 2016 : 8:30 AM to 10:20 AM
• Sponsor: Biometrics Section
• Author(s): Ross L. Prentice* and Shanshan Zhao
• Companies: Fred Hutchinson Cancer Research Center and National Institute of Environmental Health Sciences
• Keywords: Hazard ratio ; Cross ratio ; Regression ; Semiparametric ; Maximum likelihood

Abstract

Semiparametric regression models of multiplicative form are specified for marginal hazard rate processes, and for the cross ratio process for the analysis of absolutely continuous censored bivariate failure time data. The cross ratio process at follow-up time is defined as the ratio of the double failure hazard rate at (t1,t2) to the corresponding product of single failure hazard rates.These models determine a semiparametric likelihood function which can be maximized over baseline marginal hazard rate and cross ratio functions to give a profile likelihood for all regression parameters, that characterizes the dependence of the cross ratio process on regression variables, and also has potential to enhance the efficiency of marginal hazard ratio parameter estimates. Simulation evaluations will be presented, along with an application to postmenopausal hormone therapy in relation to coronaryheart disease and stroke.